23 okt 2012

Framing the crisis: accident, fiasco or debacle

The New York City headquarters of Lehman Brothers

Engelen and colleagues discuss how to understand the current crisis: is it an accident, a fiasco or a debacle?

There are only two stories: 'We name the guilty man' and 'Arrow points to defective part'. Everything else is PSJ—public service journalism. (Murray Sayle, quoted in an obituary in The Times, 21 September 2010)

Our intention was to write a book covering the financial crisis that began in 2007 in several high-income countries, concentrating on the United Kingdom but also covering the United States and the European Union. The aim was to answer two key questions in an accessible way which could influence social scientists and the political classes: how did finance come to cause the crisis, and why is it now so difficult to manage the consequences and to reform finance? As the chapters were drafted, we were increasingly troubled by a set of prior issues about how others were framing the crisis and how we should frame the crisis. Framing was relevant because it raised issues which, in academic terms, were about agency versus structure and, in popular terms, were about who or what to blame. Should we understand the crisis as an unfortunate accident caused by some kind of defect or mistake in a complex system; or did the crisis involve culpable irresponsibility or misjudgement by groups like investment bankers or regulators? Not least, framing matters because these issues are of broad interest and relevance for anyone interested in understanding present-day capitalism.

According to journalistic aphorism, there are only two stories: 'arrow indicates defective part' or 'we name the guilty men'. In a Times obituary, this was attributed to Murray Sayle, who also allegedly dismissed everything else as boring but worthy public service journalism. In other versions of the aphorism, Sayle, who was a great contrarian on issues like Bloody Sunday or the dropping of the first atom bomb, concedes there is a third kind of revisionist story because 'everything you thought you knew about this subject is wrong'. Our natural inclination was towards the third kind of story and we hoped to produce the academic equivalent of public service journalism with some revisionism. But, as we drafted chapters about financial innovation and the politics of reform, we were distracted by simultaneously reading the burgeoning academic and media literatures on crisis, which generally defaulted onto the two other stories by identifying the defective part or naming the guilty men.

Most notably, the elite British intelligentsia, including critical and independent key figures like the historian Donald MacKenzie, the regulator Andrew Haldane, and the Financial Times journalist Gillian Tett, was, by 2010, adopting variants on an accident account which associated financial crisis with disasters like the Challenger Space Shuttle or Deep Water Horizon. This framing had been adumbrated in the earlier emphasis on global imbalances in the work of figures like Adair Turner or Martin Wolf. But the critical intelligentsia added a new emphasis on mathematization and the performativity of formal knowledges. If the crisis was not an accident, our research also suggested that the crisis was not a fiasco in the classical sense familiar from the older social science literature by international authors like Bovens and 't Hart (1996) or Scott (1998); nor did we wish to endorse the attempts of those like Charles Perrow (2010) who blame the guilty bankers and politicians.

After some reflection, it seemed to us that the financial crisis could best be understood in a new and different frame as an elite debacle, which associated it with failed interventions like the American and British military ventures in Iraq and Afghanistan. This reframing was supported by research that indicated the importance of informal knowledges and of bricolage practices inside finance, which together open up new possibilities of attributing responsibility without scapegoating the guilty men; it echoes an original meaning of debacle as a confused rout. More broadly, the reframing of the financial crisis as debacle is also helpful because it situates the financial crisis in political terms as part of a much larger current problem about how and why the democratic system of political competition is not working to articulate alternatives and solutions.

The aim of this preface is then to present the book as an intervention in the ongoing debate about the crisis as accident. Through argument and evidence in the eight chapters we aim to persuade our readers to take our debacle framing seriously. Realistically, we do not aim to succeed by establishing a new mainstream orthodoxy but rather by provoking and persuading diverse readers to apply the debacle framing to their own work.

The a priori of accident

The definition and usage of the term 'systems accident' highlights the political and moral ambiguity of this kind of framing. Systems accident was originally used as a forensic category to describe an 'unanticipated interaction of multiple failures' in a complex system which is interactive and tightly coupled (Perrow 1984). The concept then provided the frame for classic accident case studies of engineering and process control disasters like the 1986 Challenger Space Shuttle disaster or the 1979 Three Mile Island nuclear accident. But, in current usage, systems accident has become the stock excuse of practitioners and corporate elites after things have gone disastrously wrong (Moran 2001). Thus, Tony Hayward, then chief executive of BP, when questioned in a Congressional hearing, described the oil spill from the Deepwater Horizon rig as 'a complex accident caused by an unprecedented combination of failures in a number of different, related processes, systems and equipment' (Plungis and Snyder 2010).

There is a large slippage between the original, austere forensic usage of the concept and the current, apologetic usage. For Perrow, the problem is about unanticipated or unforeseen, unknowable interactions between various components or events, under conditions of complexity and tight coupling. In his classic work, he was never a technocratic optimist who believed in fixing the defective part. Perrow argued more radically that complex systems which are tightly coupled should not be built because accident is inevitable (even though its precise form cannot be predicted ex ante). In the apologetic usage, the system has been built and failed disastrously and the corporate operator is trying to manage blame around the idea that accidents will happen. Thus, systems accident is not so much one explanation as an opening onto a field of accident explanation within systems.

But classic accident explanations typically share an a priori which is necessary to forms of explanation where 'arrow indicates defective part'. The system has one clear objective (such as oil from deep sea, electricity from nuclear energy, or astronauts into space) so that success means safe and efficient performance of function. The accident involves a sequence of events and failures which go critical because of a decisive technical miscalculation or defective part (often organizationally embedded): at Three Mile Island, for example, the problem was the failure to recognize coolant loss consequent upon a stuck valve; with Challenger, the problem was poor design of the O ring component, compounded by low temperature after frost on the night before launch. There is an underlying fatalism about the past because accidents will happen in a complex world. Systems accidents are sometimes described as 'normal accidents' because under some combination of technological and social conditions we must expect catastrophic outcomes. But there is also technocratic optimism about the future because relevant interactions can be mapped and analysed; and on that basis future accidents can be prevented.

In earlier UK crises, like the Barings collapse of the mid-1990s or the secondary banking crisis of the mid-1970s, complacent practitioners had dared to try the 'accidents will happen' excuse (see Moran 1986, 2001). The 'disaster' metaphor has been much used in US media reporting of the crisis, which of course does suggest the crisis is, like a hurricane, an act of God for which no one is to blame. It is perhaps more surprising to see critical and independent elite members of the British academic, regulatory, and media intelligentsia all now presenting different accounts of the financial crisis as an accident within a system because the (rectifiable) problem is with systems, not actors. Thus, accident is invoked by Britain's most distinguished contemporary historian of finance, Donald MacKenzie. Following a classic pre-crisis study of the performativity (and counter-performativity) of mainstream finance, after the crisis, MacKenzie focuses on the role of default correlation assumptions.

Accident is also reinterpreted by the most intellectually radical of our current regulators, Andrew Haldane, financial stability director of the Bank of England. The Bank's house intellectual has lost faith in mainstream economics but favours a new, biological understanding of the financial crisis as ecological or epidemiological network accidents. In the media, the anthropologically trained business journalist Gillian Tett of the Financial Times, in her latest work on the crisis, produces another new account of the accident emphasizing the problem of fragmentation of understanding which is the consequence of technocratic elites acting in silos (Tett 2010).

In MacKenzie's account (2010) of the crisis, default correlation calculations within the ABS CDO (asset-backed securities collateralized debt obligation) class of credit derivatives has much the same significance as the mis-engineering of the O ring on the Challenger Space Shuttle. His working paper reports on a case study of 'evaluation practices' in complex instruments based on seventy-six interviews, focusing on the rise of a new derivative instrument ABS CDOs, or CDOs, whose assets were ABSs or residential mortgage-backed securities. The issued value of this class of derivative ballooned to $308 billion in 2006, mainly from pools of sub-prime debt. Different teams and valuation practices had previously been applied to estimate the risk of default on CDOs related to corporate bonds, and the risk of prepayment on ABSs related to mortgages. The problem of valuing ABS CDOs was solved by using existing corporate CDO models and borrowing their correlation values for the probability that different households within a diversified mortgage pool default simultaneously. The 0.3 value, lifted from experience of corporate bond cross default, both made ABS CDOs profitable and, within MacKenzie's performativity frame, resulted in the extension of mortgages to riskier households. In doing so, this brought into being a reality that did not conform with the underlying expectations of the model. Accident is explicitly invoked because the fatal miscalculation about default correlation resulted from 'two institutionally separate insights' (MacKenzie 2010: 79). The problem was not greed for fees but was instead 'reminiscent of the rigidities and barriers to information flow in the background of the Challenger disaster' (2010: 77).

Haldane is less socially constructionist about performativity but considerably more intellectually radical than MacKenzie about the uselessness of mainstream economics, the need to rethink the crisis as stress in a complex system, and to invent a new practice of macro-prudential regulation. Haldane's key paper on 'Rethinking the financial network' (2009a) proposes a move from physics-based concepts of economics to biology-based concepts with a new, epidemiological and ecological understanding. The financial crisis represents the behaviour under stress of 'a complex, adaptive system' on the model of the spread of SARS and HIV, or the collapse of fish stocks. There is an isomorphism about 'seizures in the electrical grid, degradation of eco-systems, the spread of epidemics and the disintegration of the financial system' (Haldane 2009a: 3). The explanation is that robust but fragile networks are 'accidents waiting to happen', so that modest events can precipitate a tipping point which will be made worse by homogeneous monoculture or hideand-flight responses, which have their analogues in financial markets prone to illiquidity and dumping assets. Haldane is optimistic about the possibility of a technocratic fix for finance which would create a natural order with greater stability and resilience: this requires a new project to 'map the global financial system' and then 'vaccinate the super spreaders' (2009a: 24) or high risk, high infection individuals, and/or to institute 'central counterparties' (2009a: 29).

As an academically trained anthropologist, Gillian Tett has a rather different, more cultural take on accident, where actors play a larger role but systems limit information flow and understanding. Her popular book, Fool's Gold, told the story of the invention of derivatives by JP Morgan bankers and their subsequent diffusion, but did not turn the crisis itself into a coherent story. This task is now taken up in Tett's contribution (2010) to the Banque de France Financial Stability Review, which centres analysis on the problem of technocratic elites in their silos. There is an endemic twenty-first century problem about 'mental and structural fragmentation' in an increasingly interconnected world which helps to explain the disasters of complex credit or BP's oil spill in the Gulf. Insiders and outsiders alike could not 'join up the dots and see how systemic risks were building up in the (financial) system' (Tett 2010: 129). More precisely, there are two interrelated obstacles to understanding. First, there is a problem about technocratic elites operating in silos which are both structural, arising from the organization of banking and regulation, and cognitive, arising from how bankers and financiers conceive of finance. The second set of problems arises from Bourdieusian social 'silence': many topics, like derivatives before 2007, are not publicly discussed because they are thought boring, arcane, taboo, or unthinkable. Tett's culturally inflected fix is 'more holistic modes of thought' (2010: 129) via the employment of cultural intermediaries with an anthropological sensibility who can explain practices and mediate understandings of different worlds.

One of the peculiarities of all three accounts by MacKenzie, Haldane, and Tett is their weak visualization because there is no diagram of the accident. Classical systems accident analysis is usually supported by a process flow diagram: a sequence of malfunction, mis-steps, unanticipated and unregistered consequences produce a standard diagram of disaster in official reports and newspapers. But MacKenzie and Haldane provide no such diagram, and indeed their analyses in different ways all make the systems unnecessary or, as yet, unavailable. MacKenzie is preoccupied with transformation steps (not longer chains or circuits), as with his figures that show the ABS or cash CDO with pooled assets becoming tranched securities, or the transformation of 'mortgage backed securities into ABS CDOs' (2010: 107). Haldane (2009a) sidesteps process diagrams by identifying the need for (but not providing) a new macro map of the (whole) financial system. Tett sets up a related task because her problem of knowledge is actors but without a diagram that 'joins the dots'; like Haldane, her objective is greater legibility.

Yet, all three authors make strong assumptions about a world of expertise. This centres on the role of formal knowledge as a camera or engine, including of course the possibilities of the wrong lens or a misfire. MacKenzie's cumulative work on finance provides a history of mathematization which, in Tett's story, is what makes finance arcane; while Haldane proposes a re-mathematization of the world. There is no analysis of informal rhetorics, or how, for example, impossibilist ideas like shareholder value change the world. Neither is there reference to alternative (non-mainstream) economic paradigms: behavioural finance is not explored and the heterodox macroeconomics of the post-Keynesians and Minskians is ignored, even though the latter anticipated instability from finance. If we exclude Tett, whose mission is to persuade power to recognize its limits, there is little analysis of power and authority behind doxa: the heterodox have, in effect, over the past thirty years been purged from the academic communities which MacKenzie studies and Haldane inhabits.

Not a fiasco

Can the crisis be related to another set of policy literatures about fiascos? A fiasco was defined by Bovens and 't Hart (1996: 215) as '(i) [a] negative event that is (ii) perceived by a socially and politically significant group of people to be at least partially caused by (iii) avoidable and (iv) blamable failures of agents . . . '. These two authors put the primary emphasis on perceptions in constructionist studies (e.g. Bovens and 't Hart 1996; Bovens et al. 2001), where the task is not to explain fiasco but to explore the different meanings we give to fiasco. This is of limited relevance to our argument because we are fairly sure that the financial crisis will not turn out like the Sydney Opera House, which began as a fiasco and ended as a triumphant icon (Dunleavy 1995). But there is another literature on fiasco, represented by Scott (1998), which is much more relevant to our purposes and avoids crude scapegoating of villains like bankers. The problem, as we will argue below, is that Scott's account of crisis presents modernist governmentality as the central knowing subject of a unitary historical process that always fails in the same way.

Scott's classic study (1998) presents an anti-modernist account of 'how certain schemes to improve the human condition have failed'. Fiascos are typically the result of what (after Foucault) we would now call governmentality, operating in a particular historical conjuncture. The focus is on a toxic combination of modern state power and the Enlightenment legacy of an obsession with legibility, simplification, and measurement. The result is high modernism as an ideology, which is shaping arenas as diverse as the modern city, economic planning, and the management of nature. 'Thin simplification' – knowledge derived from standardized measurement systems – overrides métis, the practical knowledge derived from everyday experience, with catastrophic results. This is interestingly anti-modernist because Scott's verdict echoes Oakeshott's argument (1962) against rationalism in politics and for the primacy of tacit knowledge based on elite experience over expertise and data in the practice of government.

There is much to be said for Scott's account, for it does help us to understand the Reagan and Thatcher projects that combined rhetoric with design for the reconstruction of political and economic institutions after 1979. What is crucial in Scott's view is the extent to which a project reflects the attempt to make something legible: standardization; the dominance of formal, official knowledge; and the performative use of state power together transform a multidimensional reality into something that closely resembles the maps, models, and images of the world used as norm by the elites pushing for greater legibility. As Moran (2007) has argued, in a paradoxical way, Thatcher and Reagan's attempt to transform society into a market place very much fits the template of high modernist socialist projects that Scott describes. In effect, the 'neo-liberal' project is the use of state power to remake a market society according to the image of the market propounded by mainstream economics. If the neo-liberal project has the same instruments, ethos, and epistemology, but different aims, this is a high modernist project even though what is being made is the opposite of Le Corbusier's 'machines for living'.

But this general position does not deal with the disconnects, anomalies, and contradictions so ubiquitous in present-day capitalism. We have serious doubts as to whether the neo-liberal agenda was operated or operable in every important area of policy. Our research has highlighted the many ways in which developments in the finance sector after the 1980s deregulation do not fit with Scott's assumptions. As we argue in Chapter 1 on the 'great complacence', the governmental approach pre-2007 in all major jurisdictions was the very opposite of an obsessive modern concern with control, monitoring, and surveillance at the expense of métis. Key regulators, like the Financial Services Authority (FSA) in the United Kingdom, did not pursue the legibility project, through a mixture of deference to market actors, passivity, and sheer incompetence. More generally, the evidence is that much of the failure can be laid at the door of policies that dismantled monitoring and control in the name of deregulation, placed excessive faith in market operators, and placed too heavy a reliance on the tacit, practical knowledge of those with expertise in markets. It was deference to métis, not its extinction, that helped create the crisis as policy elites bought into notions of market omniscience.

If Scott's notion of fiasco will not work for deregulated finance, we cannot then take an intellectual shortcut to explanation by naming the guilty men and scapegoating the legislators, regulators, and bankers who built and operated deregulated finance. This is the explanatory strategy of Perrow (2010) for whom, in effect, a kind of political financial complex has in the present day replaced the military industrial complex which Wright Mills (1956) analysed some fifty years ago. Perrow's denial (2010) that the financial crisis is a systems accident is useful and authoritative because Perrow originated that concept. But Perrow then immediately defaults onto naming the guilty men who turn out to be senior bankers and politicians who knew what they were doing. Elites of 'key agents who were aware of the great risks' and 'crafted the ideologies and changed institutions, fully aware that this could harm their firms, clients and the public' (2010: 309).

The problem is that Perrow's judgement is, in Scottish legal terminology, 'not proven'. Perrow observes a pattern of corporate donations by financial firms and of elite ties through the revolving door between finance and politics, but that indirect evidence does not prove that money or connections always or usually suborn the independent judgement of politicians and regulators. Furthermore, it is simply not proven to argue that senior bankers all shared the same cynical understanding of those like the former Goldman Sachs CDO trader 'Fabulous Fab' Fabrice Tourre (Jenkins and Guerrera 2010). Nor is it plausible to argue that they all or mostly knew what they did and understood the consequences of their actions if, as we argue in Chapter 2, the financial innovators were bricoleurs creating a changing latticework of circuits which neither practitioners nor regulators understood. As for warnings being ignored, in Chapter 1 we demonstrate that those in authority positions in central banking and regulation were all confidently supportive of financial innovation; while the warnings from the post-Keynesians and Minskians were general ones about the unsustainable housing bubble, not specific ones about how shadow banking would blow up the world.

So the question arising from our research (and our reservations about Scott and Perrow's explanations) is: can we have a non-accidental explanation which both recognizes the agency of bankers and regulators and assigns responsibility without supposing that they fully know what they do? If so, we would have new insights into the crisis.

Elite debacle and hubris

Our research has convinced us that investment bankers, regulators, and the political classes had different kinds of agency. Politicians and the media find it easier to make jibes about investment bankers both because they were managerially in charge and because failed senior bankers like Fred Goodwin or Dick Fuld behaved publicly in such a graceless way. But our research on financial innovation as bricolage in Chapter 2 suggests that no banking insider from one node had an overview of the changing latticework of circuits.

From this point of view, the crisis resulted from an accumulation of small, and in themselves relatively harmless, decisions made by individual traders or bankers and banks. It is hard to be so kind about the regulators and the political elite who made and implemented policy in finance. They typically bought into the high modernist macro project of 'perfecting the market' and at the sectoral level bought into a 'trust the bankers to deliver functioning markets' story. This promised everything and offered very little except the undermining of public regulation, while innovation delivered the exact opposite of the promises, as risk was concentrated not dispersed by a dysfunctional banking system. In our view, this complacence was an elite debacle.

When considering Anglo-American political and economic elites, we need to distinguish between the 1980s commitment to a project of social reconstruction in the image of a deregulated system of free market capitalism and the early 2000s complacency about financial innovation in the middle of a bubble which was misread as the Great Moderation or, in Chancellor Gordon Brown's words, the end of boom and bust. The 1980s position might have been naive, but the later 2000s position was certainly hubris in the more or less exact meaning of that word: an overbearing self-confidence that led to ruin. Among leaders of institutions it is, as Owen (2007) has explored, an occupational trait: the over-confident are attracted to leadership; and once in command, especially of well-resourced institutions like modern states, they are encouraged to concentrate on big picture 'strategy', leaving tedious evidence and detail to subordinate technicians. Their role is to exercise judgement in a world where uncertainty means that mundane evidence alone cannot guide choice.

If we are considering debacles and hubris, the point of comparison is not accidents like Challenger or Three Mile Island but foreign military adventures like Suez, Vietnam, Afghanistan, and Iraq, which start from hubris and end in debacle: that is, humiliating failure or collapse into defeat. The a priori of debacle is very different from that of accident in three ways. First, informal knowledges are central to elite (mis)calculation. This can take the form of overestimating the enemy, as when Bush and Blair argued that intelligence showed Saddam Hussain had weapons of mass destruction. Or it can take the form of underestimation, as with the French commander who did not believe that the Vietcong could bring up artillery and bombard Dien Ben Phu.

Second, intervention usually has multiple, fantastic, and contradictory objectives, which are often disengaged from operating detail. The Iraq intervention was in this respect classic because it mixed high, and probably unattainable, aims of nation building and democratization with real politik about stabilizing the region, building a bulwark against Iran (and maybe controlling oil resources too).

Third, matters are greatly complicated by the unforeseen improvisation of local players in response to events and happenstance decisions by the major power. In the Iraq case, the key decision was that of Bremner and Rumsfeld when they disbanded the Iraqi army and thereby empowered every militant or gangster with a gun. In Afghanistan, the British deployment of lightly armoured vehicles used in Northern Ireland encouraged the use of roadside improvised explosive devices.

Thus, debacle is unlike an accident because the outcome is not reversible or fixable, nor even avoidable next time. A debacle which ends in defeat or withdrawal leads to shifts in the power balance; for the major power defeat is dangerous because it often (but not always) discredits the elites who embarked on the adventure.

The role of hubris in modern debacles is now being most closely documented in studies of the Iraq and Afghanistan conflicts – the former a subject also of Owen's study (1997). Afghanistan and, especially, Iraq arose from a powerful tradition in British policymaking: the belief that Britain has a distinct providential mission to export values and institutions to foreign places, and to reshape civil society to accommodate those exported values and institutions. In the nineteenth and early twentieth centuries, this providentialism took the form of imperialism and of Christian missionary projects: two fine accounts are Colley (1992) and Cannadine (2002). The Iraq invasion is the greatest foreign policy fiasco for at least seventy years. It dwarfs even the Suez disaster of 1956, since its duration, damage, and magnitude have been much greater – especially in the suffering inflicted on the people of Iraq. Over 100,000 Iraqis have died as a direct result of the invasion, which ended with the humiliating withdrawal of US and UK forces as the dysfunctional democracy is on the point of being drawn in to the Iranian field of influence. It is only comparable to the succession of military disasters which led to the fall of the Chamberlain Government in 1940.

The public case for intervention in Iraq involved the manipulation of intelligence evidence assembled in secrecy; the decision processes which led to intervention, as detailed by Lord Butler's inquiry, show a pattern of casualness and informality so characteristic of the hubristic detachment from detail. Butler's verdict, though couched in the malicious understatement of a mandarinate taking its revenge on Blair, is nevertheless devastating. The charge is that the 'sofa government' reduced the scope for collective judgement.

we are concerned that the informality and circumscribed character of the Government's procedures which we saw in the context of policy-making towards Iraq risks reducing the scope for informed collective political judgment. (Butler 2004: para 611)

What was going on here is graphically illustrated by Blair's own evidence to the (at the time of writing, ongoing) Iraq inquiry chaired by Lord Chilcott. His picture of the process that led to the decision to support an Iraq invasion is a perfect example of leadership hubris as he cites evidences and arguments that were available to every newspaper editorialist. Blair's defence is that a leader has the duty – and the capacity – to make an individual judgement in a risky world:

As I sometimes say to people, this isn't about a lie or a conspiracy or a deceit or a deception, it is a decision, and the decision I had to take was, given Saddam's history, given his use of chemical weapons, given the over 1 million people whose deaths he had caused, given ten years of breaking UN Resolutions, could we take the risk of this man reconstituting his weapons programmes, or is that a risk it would be irresponsible to take? I formed the judgment, and it is a judgment in the end. It is a decision. I had to take the decision, and I believed, and in the end so did the Cabinet, so did Parliament incidentally, that we were right not to run that risk, but you are completely right, in the end, what this is all about are the risks. (Blair (2010) Oral evidence to the Iraq Inquiry, italics added)

In our view, reframing the crisis as debacle and focusing on hubris allows us to present much the most compelling account of the ongoing financial crisis. Finance is not only an economically unsafe and violently pro-cyclical sector but also part of a democracy that is not working. 

This is the introduction to the upcoming book After the Great Complacence by Engelen et al. These are the uncorrected author proofs, pre-published with permission of the author.

Hoogleraar Financiële Geografie aan de UvA.
Alle artikelen